Opinion and Relationship Mining in Online Forums

Abstract

The Internet contains an increasing number of online forums where consumers exchange product opinions. It is important for companies to know how consumers judge their products and how these opinions are spread by interactions throughout online forums. With this knowledge it is possible to recognize risks and chances. However, classical opinion research is very time consuming and only possible to a certain extent. This paper introduces a system which automatically extracts opinions and communication relationships in forums by text mining and identifies influential users and trends by social network analysis.